# HybridVLA融合扩散与自回归，机器人操作成功率提升19%

> 原文链接：https://robotics.hellobike.com/papers/2025-03-13-2503-10631-9e2c824c
> 频道：papers | 更新日期：2025-03-13 | 标签：auto, arxiv, cs.RO, landmark
> 核心摘要：针对VLA模型动作预测离散化或推理割裂痛点，提出统一框架HybridVLA。通过协同训练将扩散去噪融入Next-token预测，并设计自适应集成机制，在仿真与真机任务中平均成功率分别提升14%和19%，实现未见配置下的稳定操作。

# HybridVLA融合扩散与自回归，机器人操作成功率提升19%

::: v-pre
> 本文由 AI 根据公开内容整理，可能有误，请以原文为准：[HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model](https://arxiv.org/abs/2503.10631)（arXiv 2503.10631）

## 核心内容

针对现有视觉-语言-动作（VLA）模型在动作连续性控制与语义推理能力上的割裂问题，研究团队提出HybridVLA统一框架。该方法旨在解决自回归方法因动作离散量化导致的精度损失，以及纯扩散方法未能充分利用大模型Token级推理能力的局限。实验数据显示，HybridVLA在仿真环境及真实世界机器人操作任务中的平均成功率较此前SOTA方法分别提升14%和19%，并在未见过的场景配置下展现出稳定的泛化操控能力。

技术架构上，HybridVLA在单一LLM内融合了扩散模型的连续动作生成特性与自回归的上下文推理能力。为消除两种生成范式的干扰，论文提出协同训练策略，将扩散去噪过程无缝嵌入Next-token预测流程，使两者相互增强。此外，系统设计了协同动作集成机制，鉴于两种方法在不同任务中表现各异，该机制能自适应融合双重预测结果，从而在保证语义理解的同时实现高精度连续控制，显著提升复杂动态环境下的鲁棒性。

## 关键看点

- **统一架构**：在单一LLM内融合扩散连续动作生成与自回归上下文推理，解决离散量化精度损失与推理割裂问题
- **协同训练**：将扩散去噪嵌入Next-token预测流程，消除范式干扰，使连续控制与语义推理能力相互增强
- **自适应集成**：设计协同动作集成机制，根据任务特性动态融合双重预测结果，提升复杂场景操控鲁棒性
- **性能验证**：仿真与真机任务平均成功率较SOTA分别提升14%和19%，在未见配置下保持稳定操作泛化能力

## 工程与产业观察

HybridVLA验证了在统一模型内协同训练扩散与自回归范式的有效性，解决了具身智能中高层语义推理与底层连续控制的耦合难题。其自适应集成机制为多模态大模型落地精密操作提供了新路径，表明未来VLA研发应重点关注异构生成头在Token级的深度融合而非简单拼接，这对提升机器人非结构化环境作业能力具有关键工程参考价值。

## 英文摘要 (Abstract)

A fundamental objective of manipulation policy design is to endow robots to comprehend human instructions, reason about scene cues, and execute generalized actions in dynamic environments. Recent autoregressive vision-language-action (VLA) methods inherit common-sense reasoning capabilities from vision-language models (VLMs) for next action-token prediction. However, these methods quantize actions into discrete bins, which disrupts the continuity required for precise control. In contrast, existing diffusion-based VLA methods incorporate an additional diffusion head to predict continuous actions solely conditioned on feature representations extracted by the VLM, without fully leveraging the VLM's pretrained reasoning capabilities through token-level generation. To address these limitations, we introduce HybridVLA, a unified framework that absorbs the continuous nature of diffusion-based actions and the contextual reasoning of autoregression within a single large language model. To mitigate interference between the two generation paradigms, we propose a collaborative training recipe that seamlessly incorporates diffusion denoising into the next-token prediction process. With this recipe, we find these two action prediction methods not only reinforce each other but also exhibit varying strength across different tasks. Therefore, we design a collaborative action ensemble mechanism that adaptively fuses both predictions, leading to more robust control. HybridVLA outperforms previous state-of-the-art VLA methods by 14\% and 19\% in mean success rate on simulation and real-world tasks, respectively, while demonstrating stable manipulation in unseen configurations.

作者：Jiaming Liu, Hao Chen, Pengju An, Zhuoyang Liu, Renrui Zhang, Chenyang Gu, Xiaoqi Li, Ziyu Guo, Sixiang Chen, Mengzhen Liu, Chengkai Hou, Mengdi Zhao, KC alex Zhou, Pheng-Ann Heng, Shanghang Zhang
:::

## 原始参考来源 (Sources)

- [arXiv 2503.10631：HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model](https://arxiv.org/abs/2503.10631)